ipv4: lock mtu in fnhe when received PMTU < net.ipv4.route.min_pmtu

Prior to the rework of PMTU information storage in commit
2c8cec5 ("ipv4: Cache learned PMTU information in inetpeer."),
when a PMTU event advertising a PMTU smaller than
net.ipv4.route.min_pmtu was received, we would disable setting the DF
flag on packets by locking the MTU metric, and set the PMTU to
net.ipv4.route.min_pmtu.

Since then, we don't disable DF, and set PMTU to
net.ipv4.route.min_pmtu, so the intermediate router that has this link
with a small MTU will have to drop the packets.

This patch reestablishes pre-2.6.39 behavior by splitting
rtable->rt_pmtu into a bitfield with rt_mtu_locked and rt_pmtu.
rt_mtu_locked indicates that we shouldn't set the DF bit on that path,
and is checked in ip_dont_fragment().

One possible workaround is to set net.ipv4.route.min_pmtu to a value low
enough to accommodate the lowest MTU encountered.

Fixes: 2c8cec5 ("ipv4: Cache learned PMTU information in inetpeer.")
Signed-off-by: Sabrina Dubroca <[email protected]>
Reviewed-by: Stefano Brivio <[email protected]>
Signed-off-by: David S. Miller <[email protected]>

Author: qsn

include/net/ip.h

@@ -328,14 +328,21 @@ int ip_decrease_ttl(struct iphdr *iph)
 	return --iph->ttl;
 }
 
+static inline int ip_mtu_locked(const struct dst_entry *dst)
+{
+	const struct rtable *rt = (const struct rtable *)dst;
+
+	return rt->rt_mtu_locked || dst_metric_locked(dst, RTAX_MTU);
+}
+
 static inline
 int ip_dont_fragment(const struct sock *sk, const struct dst_entry *dst)
 {
 	u8 pmtudisc = READ_ONCE(inet_sk(sk)->pmtudisc);
 
 	return  pmtudisc == IP_PMTUDISC_DO ||
 		(pmtudisc == IP_PMTUDISC_WANT &&
-		 !(dst_metric_locked(dst, RTAX_MTU)));
+		 !ip_mtu_locked(dst));
 }
 
 static inline bool ip_sk_accept_pmtu(const struct sock *sk)
@@ -361,7 +368,7 @@ static inline unsigned int ip_dst_mtu_maybe_forward(const struct dst_entry *dst,
 	struct net *net = dev_net(dst->dev);
 
 	if (net->ipv4.sysctl_ip_fwd_use_pmtu ||
-	    dst_metric_locked(dst, RTAX_MTU) ||
+	    ip_mtu_locked(dst) ||
 	    !forwarding)
 		return dst_mtu(dst);
 

include/net/ip_fib.h

@@ -59,6 +59,7 @@ struct fib_nh_exception {
 	int				fnhe_genid;
 	__be32				fnhe_daddr;
 	u32				fnhe_pmtu;
+	bool				fnhe_mtu_locked;
 	__be32				fnhe_gw;
 	unsigned long			fnhe_expires;
 	struct rtable __rcu		*fnhe_rth_input;

include/net/route.h

@@ -63,7 +63,8 @@ struct rtable {
 	__be32			rt_gateway;
 
 	/* Miscellaneous cached information */
-	u32			rt_pmtu;
+	u32			rt_mtu_locked:1,
+				rt_pmtu:31;
 
 	u32			rt_table_id;
 

net/ipv4/route.c

@@ -634,6 +634,7 @@ static inline u32 fnhe_hashfun(__be32 daddr)
 static void fill_route_from_fnhe(struct rtable *rt, struct fib_nh_exception *fnhe)
 {
 	rt->rt_pmtu = fnhe->fnhe_pmtu;
+	rt->rt_mtu_locked = fnhe->fnhe_mtu_locked;
 	rt->dst.expires = fnhe->fnhe_expires;
 
 	if (fnhe->fnhe_gw) {
@@ -644,7 +645,7 @@ static void fill_route_from_fnhe(struct rtable *rt, struct fib_nh_exception *fnh
 }
 
 static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw,
-				  u32 pmtu, unsigned long expires)
+				  u32 pmtu, bool lock, unsigned long expires)
 {
 	struct fnhe_hash_bucket *hash;
 	struct fib_nh_exception *fnhe;
@@ -681,8 +682,10 @@ static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw,
 			fnhe->fnhe_genid = genid;
 		if (gw)
 			fnhe->fnhe_gw = gw;
-		if (pmtu)
+		if (pmtu) {
 			fnhe->fnhe_pmtu = pmtu;
+			fnhe->fnhe_mtu_locked = lock;
+		}
 		fnhe->fnhe_expires = max(1UL, expires);
 		/* Update all cached dsts too */
 		rt = rcu_dereference(fnhe->fnhe_rth_input);
@@ -706,6 +709,7 @@ static void update_or_create_fnhe(struct fib_nh *nh, __be32 daddr, __be32 gw,
 		fnhe->fnhe_daddr = daddr;
 		fnhe->fnhe_gw = gw;
 		fnhe->fnhe_pmtu = pmtu;
+		fnhe->fnhe_mtu_locked = lock;
 		fnhe->fnhe_expires = expires;
 
 		/* Exception created; mark the cached routes for the nexthop
@@ -787,7 +791,8 @@ static void __ip_do_redirect(struct rtable *rt, struct sk_buff *skb, struct flow
 				struct fib_nh *nh = &FIB_RES_NH(res);
 
 				update_or_create_fnhe(nh, fl4->daddr, new_gw,
-						0, jiffies + ip_rt_gc_timeout);
+						0, false,
+						jiffies + ip_rt_gc_timeout);
 			}
 			if (kill_route)
 				rt->dst.obsolete = DST_OBSOLETE_KILL;
@@ -1009,15 +1014,18 @@ static void __ip_rt_update_pmtu(struct rtable *rt, struct flowi4 *fl4, u32 mtu)
 {
 	struct dst_entry *dst = &rt->dst;
 	struct fib_result res;
+	bool lock = false;
 
-	if (dst_metric_locked(dst, RTAX_MTU))
+	if (ip_mtu_locked(dst))
 		return;
 
 	if (ipv4_mtu(dst) < mtu)
 		return;
 
-	if (mtu < ip_rt_min_pmtu)
+	if (mtu < ip_rt_min_pmtu) {
+		lock = true;
 		mtu = ip_rt_min_pmtu;
+	}
 
 	if (rt->rt_pmtu == mtu &&
 	    time_before(jiffies, dst->expires - ip_rt_mtu_expires / 2))
@@ -1027,7 +1035,7 @@ static void __ip_rt_update_pmtu(struct rtable *rt, struct flowi4 *fl4, u32 mtu)
 	if (fib_lookup(dev_net(dst->dev), fl4, &res, 0) == 0) {
 		struct fib_nh *nh = &FIB_RES_NH(res);
 
-		update_or_create_fnhe(nh, fl4->daddr, 0, mtu,
+		update_or_create_fnhe(nh, fl4->daddr, 0, mtu, lock,
 				      jiffies + ip_rt_mtu_expires);
 	}
 	rcu_read_unlock();
@@ -1280,7 +1288,7 @@ static unsigned int ipv4_mtu(const struct dst_entry *dst)
 
 	mtu = READ_ONCE(dst->dev->mtu);
 
-	if (unlikely(dst_metric_locked(dst, RTAX_MTU))) {
+	if (unlikely(ip_mtu_locked(dst))) {
 		if (rt->rt_uses_gateway && mtu > 576)
 			mtu = 576;
 	}
@@ -1521,6 +1529,7 @@ struct rtable *rt_dst_alloc(struct net_device *dev,
 		rt->rt_is_input = 0;
 		rt->rt_iif = 0;
 		rt->rt_pmtu = 0;
+		rt->rt_mtu_locked = 0;
 		rt->rt_gateway = 0;
 		rt->rt_uses_gateway = 0;
 		rt->rt_table_id = 0;
@@ -2546,6 +2555,7 @@ struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_or
 		rt->rt_is_input = ort->rt_is_input;
 		rt->rt_iif = ort->rt_iif;
 		rt->rt_pmtu = ort->rt_pmtu;
+		rt->rt_mtu_locked = ort->rt_mtu_locked;
 
 		rt->rt_genid = rt_genid_ipv4(net);
 		rt->rt_flags = ort->rt_flags;
@@ -2648,6 +2658,8 @@ static int rt_fill_info(struct net *net,  __be32 dst, __be32 src, u32 table_id,
 	memcpy(metrics, dst_metrics_ptr(&rt->dst), sizeof(metrics));
 	if (rt->rt_pmtu && expires)
 		metrics[RTAX_MTU - 1] = rt->rt_pmtu;
+	if (rt->rt_mtu_locked && expires)
+		metrics[RTAX_LOCK - 1] |= BIT(RTAX_MTU);
 	if (rtnetlink_put_metrics(skb, metrics) < 0)
 		goto nla_put_failure;
 

net/ipv4/xfrm4_policy.c

@@ -100,6 +100,7 @@ static int xfrm4_fill_dst(struct xfrm_dst *xdst, struct net_device *dev,
 	xdst->u.rt.rt_gateway = rt->rt_gateway;
 	xdst->u.rt.rt_uses_gateway = rt->rt_uses_gateway;
 	xdst->u.rt.rt_pmtu = rt->rt_pmtu;
+	xdst->u.rt.rt_mtu_locked = rt->rt_mtu_locked;
 	xdst->u.rt.rt_table_id = rt->rt_table_id;
 	INIT_LIST_HEAD(&xdst->u.rt.rt_uncached);
 	rt_add_uncached_list(&xdst->u.rt);